強夯加固深度預(yù)測的神經(jīng)網(wǎng)絡(luò)法與等效擬靜力法的研究.pdf

1、天津大學(xué)碩士學(xué)位論文強夯加固深度預(yù)測的神經(jīng)網(wǎng)絡(luò)法與等效擬靜力法的研究姓名：孔慶哲申請學(xué)位級別：碩士專業(yè)：結(jié)構(gòu)工程指導(dǎo)教師：王成華李偉20021201ABSTRACTThemethodofdynamicconsolidationhasbeenincreasinglydevelopedandhasfounditsapplicationsinthefieldofground廿eatmenLsince1970swhenitwasfirstlyi

2、ntroducedintochinaduetoitsadvantagessuch船simplicityinequipmentconvenientinoperation，broadnessinserviceabilityanditseconomyincostBut，therehasnotbeenaseriesofwellestablishedmethodsforitsdesignandanalysisasitsworkingmechani

3、smofimprovementofweakgroundsisverysophisticatedTherehasbeenagreatdiversityofthedefinitionsandevaluationmethodsforeffectivedepthofimprovementbydynamicconsolidation鋤nowBasedonareviewandanalysesofresearchworksreportedinlite

4、rature，thispaperfocusesonthedevelopmentofaneuralnetworkmodelandanequivalentpseudosta垃cmethodforpredictionofimproveddepthbydynamicconsolidationTherearethreepartstotheresearchwork：First，ananalysisismadetothefactorsinfluenc

5、ingdepthofimprovementbydynamicconsolidation11lefactorsareclaSsifiedintotwocategoriesieoperationparametersinconstructiontechniquesofdynamicconsolidationendengineeringpropertiesoffoundationsoilsAneuralnetworkmodelforthepre

6、dictionoftheeffectiveDepthofimprovementissetupthroughthetrainingandexaminationpmcessesofthenetworkwithrealengineeringcaSesofdynamicconsolidationandthemodelisendedinVisualFORTRANLanguageSecond，thestate—ofthe—antofpseudost

7、aticmethodaswellasoftheftfundamentalconceptsandcalculationtechniqueswasgiven，afterwhichacomparisonoftheimproveddepthmeasuredinsomeengineeringcaSesandthosepredictedwiththeequivalentpseudostaticmethodismade，andthereasonsfo

8、rtheerlorsinthepredictionareinvestigatedAsaresultofaboveresearches，someimprovementsaremadetotheoriginalequivalentpseudostaticmethodFinally，theapplicabilityandtheeffectivenessoftheneuralnetworkmodelandthemodifiedequivalen

9、tpseudostaticmethodforpredictingtheeffectivedepthofimprovementbydynamicconsolidationaredemonstratedthroughthecomputationofanengineeringcaseinShenzhenClliuaKeywords：DynamicconsoHdation，Effectivedepthofimprovement。BPneural